• Communications for Statistical Applications and Methods
• /
• v.31 no.2
• /
• pp.235-246
• /
• 2024

In high-dimensional data analysis, sufficient dimension reduction (SDR) has been considered as an attractive tool for reducing the dimensionality of predictors while preserving regression information. The principal support vector machine (PSVM) (Li et al., 2011) offers a unified approach for both linear and nonlinear SDR. This article comprehensively explores a variety of SDR methods based on the PSVM, which we call principal machines (PM) for SDR. The PM achieves SDR by solving a sequence of convex optimizations akin to popular supervised learning methods, such as the support vector machine, logistic regression, and quantile regression, to name a few. This makes the PM straightforward to handle and extend in both theoretical and computational aspects, as we will see throughout this article.

• Environmental and Resource Economics Review
• /
• v.27 no.1
• /
• pp.195-214
• /
• 2018

This study estimates the price elasticity of residential water demand in Korea. For that, annual panel data from the year of 2010 to 2013 for 161 local water services is estimated by using panel quantile model. As a result, the price elasticities of residental water demand in Korea are estimated to be between -0.156 and -0.189 depending on its quantile. In addition, the study finds that the estimated elasticity of residential water demand by traditional conditional mean regression is relatively more influenced by high demand areas because the distribution of residental water demand in Korea is left-skewed.

• The Korean Journal of Applied Statistics
• /
• v.28 no.1
• /
• pp.83-92
• /
• 2015

Concerning the national health, it is important to develop an accurate prediction method of atmospheric particulate matter (PM) because being exposed to such fine dust can trigger not only respiratory diseases as well as dermatoses, ophthalmopathies and cardiovascular diseases. The National Institute of Environmental Research (NIER) employs a decision tree to predict bad weather days with a high PM concentration. However, the decision tree method (even with the inherent unstableness) cannot be a suitable model to predict bad weather days which represent only 4% of the entire data. In this paper, while presenting the inaccuracy and inappropriateness of the method used by the NIER, we present the utility of a new prediction model which adopts boosting with quantile loss functions. We evaluate the performance of the new method over various ${\tau}$-value's and justify the proposed method through comparison.

• The Journal of Asian Finance, Economics and Business
• /
• v.4 no.1
• /
• pp.5-18
• /
• 2017

Total debt in the People's Republic of China surged to nearly 290% as a ratio to GDP by the second quarter of 2016, mostly on account of non-financial corporate debt. The outpouring of credit to stem the impact of the global financial crisis accentuated industrial overcapacity in traditional sectors, such as steel, cement, and energy, while feeding asset bubbles in the property, equity and bond markets. At the Chinese corporate level, this has translated into weakened fundamentals and a fall in industrial profits, particularly of SOEs. As debtors struggle to service interest payments, non-performing loans (NPLs) have been on the rise. This paper assesses the financial fragility of the Chinese economy by looking at risk factors in the non-financial sector. We apply quantile regressions to a dataset containing all Chinese listed companies in Standard & Poor's IQ Capital database. We find higher sensitivity over time of corporate leverage to some of its key determinants, particularly for firms at the upper margin of the distribution. In particular, profitability increasingly acts as a curb on corporate leverage. At a time of falling profitability across the Chinese non-financial corporate sector, this eases the brake on leverage and may contribute to its continuing increase.

• Atmosphere
• /
• v.24 no.3
• /
• pp.379-390
• /
• 2014

This study explores the long-term trends of surface air temperatures in 11 KMA stations over the period of 1960~2012. Both linear and nonlinear trends are examined for the $95^{th}$, $50^{th}$, and $5^{th}$ percentiles of daily maximum ($T_{max}$) and minimum temperatures ($T_{min}$) by using quantile regression method. It is found that in most stations linear trends of $T_{max}$ and $T_{min}$ are generally stronger in winter than in summer, and warming trend of the $5^{th}$ percentile temperature (cold extreme) is stronger than that of the $95^{th}$ percentile temperature (warm extreme) in both seasons. The nonlinear trends, which are evaluated by the second order polynomial fitting, show a strong nonlinearity in winter. Specifically, winter temperatures have increased until 2000s but slightly decreased afterward in all percentiles. This contrasts with the $95^{th}$ and $50^{th}$ percentiles of summer $T_{min}$ that show a decreasing trend until 1980s then an increasing trend. While this result is consistent with a seasonal dependence of the recent global warming hiatus, most of the nonlinear trends are statistically insignificant, making a quantitative attribution of nonlinear temperature trends challenging.

• Journal of the Korean Geotechnical Society
• /
• v.30 no.12
• /
• pp.5-14
• /
• 2014

Most of slope failures have taken place between June and September in Korea, which cause a considerable damage to society. Rainfall intensity and duration are very significant triggering factors for landslide. In this paper, landslide-triggering rainfall threshold consisting of rainfall intensity-duration (I-D) was proposed. For this study, total 255 landslides were collected in landslide inventory during 1999 to 2012 from NDMI (National Disaster Management Institute), various reports, newspapers and field survey. And most of the required rainfall data were collected from KMA (Korea Meteorological Administration). The collected landslides were classified into three categories: debris flow, shallow landslide and unconfirmed. A rainfall threshold was proposed based on landslide type using statistical method such as quantile-regression method. Its validation was carried out based on 2013 landslide database. The proposed rainfall threshold was also compared with previous rainfall thresholds. The proposed landslide-triggering rainfall thresholds could be used in landslide early warning system in Korea.

• Journal of IKEEE
• /
• v.5 no.1 s.8
• /
• pp.1-7
• /
• 2001

In this paper, we have investigated the CDMA(Code Division Multiple Access) Cellular System with non-linear equalizer in reverse link channel. In general, due to unknown characteristics of channel in the wireless communication, the distribution of the observables cannot be specified by a finite set of parameters; instead, we partitioned the m-dimensional sample space Into a finite number of disjointed regions by using quantiles and a vector quantizer based on training samples. The algorithm proposed is based on a piecewise approximation to regression function based on quantiles and conditional partition moments which are estimated by Robbins Monro Stochastic Approximation (RMSA) algorithm. The resulting equalizers and detectors are robust in the sense that they are insensitive to variations in noise distributions. The main idea is that the robust equalizers and robust partition detectors yield better performance in equiprobably partitioned subspace of observations than the conventional equalizer in unpartitioned observation space under any condition. And also, we apply this idea to the CDMA system and analyze the BER performance.

• Review of Culture and Economy
• /
• v.21 no.2
• /
• pp.3-28
• /
• 2018

This study focuses on analyzing the glass ceiling effect in arts and culture professionals through the quintile decomposition applied to the RIF unconditional quantile regression and Oaxaca-Blinder decomposition technique. From the industrial viewpoint, we divide arts and culture professionals into cultural contents professionals(large category J industry) and arts professionals(large category R industry). For our analysis, we employ the pooling data of 'Wage Structure Survey' from 2009 to 2016. Our results are summarized as follows. First, as OLS wage decomposition showed that the gender wage gap among the arts professionals was lower than cultural contents professionals, but the discrimination portion of total gender wage gap was larger. Second, from quintile regression decompositions, the glass ceiling effects of two types of professionals showed different results. Cultural contents sector was observed with the "steady glass ceiling effect" as the portion of the discrimination was continuously increased, while the arts sector was observed with the "limited glass ceiling effect" as the discrimination had drastically increased in the 80s and 90s.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2018.05a
• /
• pp.23-23
• /
• 2018

본 연구에서는 다중분위회귀분석모형(Multiple Quantile Regression Model, MQRM)과 MODIS(MODerate resolution Imaging Spectroradiometer) LST (Land Surface Temperature) 자료를 이용하여 전국 공간토양수분을 산정하였다. 공간토양수분을 산정하기 위한 과정은 크게 두가지로 구분된다. 첫 번째로 기존의 MODIS LST 자료를 조건부 합성 보정기법을 적용하여 실측 LST 자료와 비교하여 위성 LST 자료가 갖고 있는 오차를 보정하였다. 그 결과, 조건부 합성 보정기법을 적용하기전 전국 71개 지상관측지점에서 관측한 실측 LST와 MODIS LST의 $R^2$는 전체 평균 0.70으로 어는정도 유의성 있는 상관관계를 나타냈으나 조건부 합성 보정기법을 적용한 후 실측 LST와 MODIS LST의 $R^2$는 전체 평균 0.92로 상당히 크게 향상됨을 알 수 있었다. 두 번째로 보정된 MODIS LST를 이용하여 다중분위회귀분석 모형을 개발하고 토양수분을 예측하는 단계로 입력자료로 위성영상 자료와 관측자료를 융합하여 사용하였다. 위성영상 자료로는 보정된 MODIS LST와 MODIS NDV를 구축하였고 일단위 강수량 및 일조시간의 기상자료는 기상청으로부터 전국 71개 지점에 대해 구축하여 IDW 공간보간기법을 이용한 공간자료로 구축하였다. 토양수분 결과를 비교하기 위한 관측 토양수분은 자동농업기상관측(Automated Agriculture Observing System, AAOS)지점에서 2013년 1월부터 2015년 12월까지의 실측 일단위 토양수분 자료를 구축하여 사용하였다. 다중분위회귀분석 모형은 LST 인자를 중심으로 각각의 분위(0.05, 0.25, 0.5, 0.75, 0.95)에 해당되는 값의 회귀식을 NDVI, 강수 입력자료를 독립인자로서 조합하여 계절 및 토성에 따른 총 80개의 회귀식을 산정하였다. 관측 토양수분과 모의 토양수분을 비교한 결과 $R^2$가 0.70 (철원), 0.90 (춘천), 0.85 (수원), 0.65 (서산), 0.78 (청주), 0.82 (전주), 0.62 (순천), 0.63 (진주), 0.78 (보성)로 높은 상관성을 보였다. 본 연구에서는 다중분위회귀 모형의 성능을 검증하기 위해 기존의 다중선형회귀모형의 결과와 비교하여 크게 개선됨을 나타냈다.

• Journal of Broadcast Engineering
• /
• v.24 no.2
• /
• pp.264-272
• /
• 2019

In this paper, we propose a 3D human body reconstruction and refinement method from the frames extracted from a video to obtain natural and smooth motion in temporal domain. Individual frames extracted from the video are fed into convolutional neural network to estimate the location of the joint and the silhouette of the human body. This is done by projecting the parameter-based 3D deformable model to 2D image and by estimating the value of the optimal parameters. If the reconstruction process for each frame is performed independently, temporal consistency of human pose and shape cannot be guaranteed, yielding an inaccurate result. To alleviate this problem, the proposed method analyzes and interpolates the principal component parameters of the 3D morphable model reconstructed from each individual frame. Experimental result shows that the erroneous frames are corrected and refined by utilizing the relation between the previous and the next frames to obtain the improved 3D human reconstruction result.